Multi-objective collaborative optimization scheduling system and method for industrial parks based on artificial intelligence

Abstract

The present invention proposes an artificial intelligence-based multi-objective collaborative optimization scheduling method and system for industrial parks, which relates to the field of intelligent scheduling technology for industrial parks. By constructing a digital twin model and a dynamic topology diagram, graph neural networks are used to model resource flow relationships, and a multi-model prediction framework is combined to perform resource demand analysis. A multi-agent reinforcement learning algorithm is used to generate scheduling strategies, and strategy distillation technology is used to achieve migration from global strategies to local decision-making modules. The robustness of the scheduling strategies is simulated and verified in the digital twin model, and the strategies are adjusted according to the simulation results, ultimately achieving efficient scheduling of park resources. This method effectively improves the intelligence and collaboration level of industrial park operations, reduces energy consumption and carbon emissions, and enhances the park's ability to cope with complex scenarios. It is suitable for the efficient management of modern industrial parks.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent scheduling of industrial parks, and more specifically, to an industrial park multi-objective collaborative optimization scheduling system and method based on artificial intelligence. Background Art

[0002] In today's context of rapid digitalization and intelligent development, the management and operation of industrial parks face many challenges. Traditional scheduling systems and methods often cannot meet the complex and changing production needs, especially in terms of resource optimization, energy management, equipment maintenance and logistics scheduling. In addition, with the continuous expansion of the scale of industrial parks and the increasing complexity of their operations, achieving multi-objective collaborative optimization has become an issue that needs to be urgently addressed.

[0003] In recent years, the rapid development of artificial intelligence technology has provided new ideas and solutions for the optimization and scheduling of industrial parks. By introducing advanced technologies such as machine learning, deep learning, graph neural networks, and multi-agent reinforcement learning, real-time collection and analysis of multi-source data such as equipment operating status, energy consumption, and logistics information in industrial parks can be achieved. For example, the use of digital twin technology to build a virtual model of the industrial park can realize dynamic simulation and optimization of resource flow relationships. At the same time, the prediction model based on artificial intelligence can accurately predict energy demand and equipment failure probability, so as to formulate reasonable scheduling strategies in advance. In addition, the multi-agent reinforcement learning algorithm can generate efficient scheduling strategies, and through centralized training and distributed execution frameworks, it can achieve an organic combination of global optimization and local decision-making. These technologies have improved the resource utilization and production efficiency of industrial parks, and significantly reduced energy consumption and operating costs.

[0004] The existing technology has the problems of low scheduling efficiency, low resource utilization and high energy consumption in industrial parks. Summary of the Invention

[0005] In order to overcome the problems of low scheduling efficiency, low resource utilization, and high energy consumption in industrial parks in the existing technology, the present invention designs an industrial park multi-objective collaborative optimization scheduling system and method based on artificial intelligence, which can effectively solve the above technical problems.

[0006] In order to solve the above technical problems, the technical solutions of the present invention are as follows:

[0007] The multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence includes the following steps:

[0008] Build a digital twin model and generate a dynamic topology map. Leverage IoT technology to collect information about the operating status of equipment, energy consumption, and logistics within the industrial park. Based on the collected data, build a digital twin model and a dynamic topology map. The nodes of the dynamic topology map include factory, warehouse, and energy station entities, and edges represent resource flow relationships between nodes.

[0009] Utilize graph neural networks for relationship modeling, define node and edge features, calculate the association weights between nodes through an attention mechanism, and use the graph neural network to learn the resource flow characteristics between nodes in the industrial park, outputting logistics path optimization solutions and energy allocation weights.

[0010] Perform resource demand analysis based on a multi-model forecasting framework, using a random forest model to predict energy demand fluctuations, a support vector machine algorithm to predict equipment failure probability, and a decision tree algorithm to classify and prioritize resource demands;

[0011] A multi-agent reinforcement learning algorithm is used to generate scheduling strategies, define a global reward function, generate joint action strategies through a centralized training-distributed execution framework, and use policy distillation technology to migrate the global strategy to the local decision-making modules of each agent.

[0012] The robustness of the scheduling strategy is simulated and verified in the digital twin model, and the strategy is adjusted based on the simulation results. The final scheduling instructions are generated and sent to the park execution terminal.

[0013] Preferably, in the step of using a graph neural network to perform relationship modeling, the nodes include equipment type, equipment status, energy consumption rate, and logistics throughput, and the edges include resource flow type, resource flow rate, and resource flow priority;

[0014] The attention mechanism determines the association weights between the nodes by calculating the Euclidean distance and cosine similarity between the nodes, and the association weights are used to adjust the edge weights of the graph neural network.

[0015] Preferably, the step of performing resource demand analysis based on a multi-model prediction framework further includes:

[0016] Use long-short-term memory networks to model the time series data of equipment operating status to predict the potential trend of equipment failure. The prediction results are then integrated with the prediction results of the support vector machine algorithm to generate a comprehensive prediction value of equipment failure probability.

[0017] The decision tree algorithm’s priority sorting logic includes: dividing resource urgency levels according to equipment failure probability thresholds, and combining them with energy demand fluctuation forecasts to generate a dynamic weight table for resource scheduling. The weight calculation method is:

[0018] W=α·P fault +β·|D predicted -D current |

[0019] Among them, α, β are adjustable parameters, P fault is the failure probability, Dpredicted and D current They are forecast demand and current demand respectively.

[0020] Preferably, generating a joint action strategy through a centralized training-distributed execution framework and migrating the global strategy to the local decision module of each agent using a policy distillation technique includes:

[0021] Define the global reward function as a weighted sum of multiple objectives:

[0022] R global =λ1·R cost +λ2·R energy +λ3·R carbon

[0023] Among them, λ1, λ2, λ3 are dynamically adjusted weight coefficients; R global is the global reward, R cost is the reward related to the cost, R energy Rewards related to energy efficiency, R carbon incentives tied to carbon emissions;

[0024] A centralized training-distributed execution framework is used to share local observation information of intelligent agents during the training phase and generate joint action strategies.

[0025] The global strategy is transferred to the local decision-making module of each agent through policy distillation technology.

[0026] Preferably, the step of simulating and verifying the robustness of the scheduling strategy in the digital twin model further includes:

[0027] Conduct random perturbation tests on the scheduling strategy using Monte Carlo simulation methods to evaluate its adaptability to different failure scenarios and demand fluctuations, and optimize and adjust the scheduling strategy based on the test results;

[0028] The step of generating the final scheduling instruction and sending it to the park execution terminal also includes:

[0029] The scheduling instructions are transmitted to the execution terminal in real time through the industrial Internet of Things platform, and the execution of the scheduling instructions is monitored and adjusted in real time through the feedback mechanism of the execution terminal.

[0030] The AI-based multi-objective collaborative optimization scheduling system for industrial parks includes:

[0031] Data collection module, used to collect equipment operating status, energy consumption and logistics information within the industrial park using IoT technology;

[0032] Digital twin construction module, used to build digital twin models and dynamic topology maps based on collected data;

[0033] The relationship modeling module is used to perform relationship modeling using graph neural networks and output logistics path optimization solutions and energy allocation weights;

[0034] Resource demand analysis module, used to perform resource demand analysis based on a multi-model prediction framework;

[0035] Scheduling strategy generation module, used to generate scheduling strategies using multi-agent reinforcement learning algorithms;

[0036] The verification and adjustment module is used to simulate and verify the robustness of the scheduling strategy in the digital twin model, adjust the strategy according to the simulation results, and generate the final scheduling instructions.

[0037] An electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence are implemented.

[0038] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-mentioned artificial intelligence-based multi-objective collaborative optimization scheduling method for industrial parks.

[0039] Compared with the existing technology, the beneficial effects of the present invention are as follows: by constructing a digital twin model and a dynamic topology map, it can reflect the operating status of the industrial park in real time and accurately, provide data support for scheduling decisions, and improve the accuracy and real-time performance of scheduling. Graph neural networks are used for relational modeling to capture the complex resource flow characteristics within the industrial park, optimize logistics paths and energy distribution, improve resource utilization efficiency, and reduce operating costs. Resource demand analysis based on a multi-model prediction framework can accurately predict energy demand fluctuations and equipment failure probabilities. Priority sorting is performed through a decision tree algorithm, ensuring the efficient allocation of key resources and emergency response capabilities. The scheduling strategy generated by a multi-agent reinforcement learning algorithm is realized through a centralized training-distributed execution framework and policy distillation technology to achieve the coordination of global strategies and local decisions, enhancing the adaptability and flexibility of the scheduling strategy. Finally, the robustness of the scheduling strategy is simulated and verified in the digital twin model, and optimized and adjusted in combination with the Monte Carlo simulation method. The present invention ensures the stability and reliability of the scheduling strategy in different scenarios. At the same time, the real-time transmission and monitoring of scheduling instructions are realized through the industrial Internet of Things platform, improving the overall operational efficiency and management level of the industrial park. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the implementation methods or the description of the prior art. Obviously, the drawings described below are merely exemplary. For ordinary technicians in this field, other implementation drawings can be derived based on the provided drawings without any creative work.

[0041] Figure 1 A step-by-step diagram of the multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence;

[0042] Figure 2 This is the structure diagram of the multi-objective collaborative optimization scheduling system for industrial parks based on artificial intelligence. DETAILED DESCRIPTION

[0043] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting this patent;

[0044] In order to better illustrate this embodiment, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product size;

[0045] It is understandable to those skilled in the art that some well-known structures and descriptions thereof may be omitted in the drawings.

[0046] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0047] Example 1

[0048] Multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence, such as Figure 1 As shown, the following steps are included:

[0049] Build a digital twin model and generate a dynamic topology map. Leverage IoT technology to collect information about the operating status of equipment, energy consumption, and logistics within the industrial park. Based on the collected data, build a digital twin model and a dynamic topology map. The nodes of the dynamic topology map include factory, warehouse, and energy station entities, and edges represent resource flow relationships between nodes.

[0050] A large number of IoT sensors are deployed in the park, including equipment operation status monitoring sensors, such as motor speed, temperature, and pressure sensors, energy consumption meters, electricity, gas, and water resource meters, as well as logistics tracking equipment, RFID tags, and GPS locators.

[0051] Sensors collect data in real time and transmit it to the park data center. Based on the collected data, the data center uses modeling software to build a digital twin model to accurately reproduce the appearance, layout and internal equipment structure of each factory, warehouse and energy station in the park. At the same time, a dynamic topology map is generated based on the actual resource flow relationship between entities. For example, factory A delivers a certain amount of finished products to warehouse B every day, and energy station C provides electricity and heat energy to factory A and warehouse B. These relationships are presented in the dynamic topology map with directed edges connecting the corresponding nodes, and the thickness of the edges can intuitively reflect the amount of resource flow.

[0052] Utilize graph neural networks for relationship modeling, define node and edge features, calculate the association weights between nodes through an attention mechanism, and use the graph neural network to learn the resource flow characteristics between nodes in the industrial park, outputting logistics path optimization solutions and energy allocation weights.

[0053] Define node and edge features. In terms of node features, equipment types include machine tools, cranes, conveyor belts, etc. Equipment status is divided into normal operation, standby, fault maintenance, etc. The energy consumption rate is calculated based on the real-time energy consumption data of different equipment, and the logistics throughput statistics the amount of goods entering and leaving each node per unit time; in terms of edge features, resource flow types include raw material transportation, finished product distribution, energy transmission, etc. The resource flow rate is calculated based on the real-time traffic data collected by the Internet of Things, and the resource flow priority is set comprehensively based on the urgency of production tasks, energy supply stability requirements, etc.

[0054] An attention mechanism is used to calculate the Euclidean distance and cosine similarity between nodes. For two adjacent factory nodes, if their equipment types are highly similar and their energy consumption rate fluctuation trends are similar, a higher association weight is obtained through complex calculations. This weight is used to dynamically adjust the edge weights of the graph neural network, so that when learning resource flow characteristics, more focus is placed on the relationship between closely related nodes. Ultimately, the optimized logistics path is output, such as planning distribution routes that avoid traffic congestion and reduce transportation distances, as well as reasonable energy allocation weights to ensure that energy is given priority supply to key production links.

[0055] Perform resource demand analysis based on a multi-model forecasting framework, using a random forest model to predict energy demand fluctuations, a support vector machine algorithm to predict equipment failure probability, and a decision tree algorithm to classify and prioritize resource demands;

[0056] Energy demand forecasting uses a random forest model, inputting historical energy consumption data, the day's weather conditions, factors affecting energy-consuming equipment such as air conditioners, factory production plans, differences in energy consumption among different products, and other multi-dimensional features to predict energy demand fluctuations in the next few hours or even days. If a factory plans to increase production of high-energy-consuming products and the temperature rises suddenly on that day, the model accurately predicts that energy demand will increase significantly.

[0057] Equipment failure prediction uses the support vector machine algorithm, combined with historical data of equipment operating parameters and maintenance records, to predict the probability of equipment failure. At the same time, the long short-term memory network is used to model the time series data of equipment operating status to capture the trend of equipment performance degradation. For example, if the vibration amplitude of a key machine tool has gradually increased and the temperature has continued to rise recently, the long short-term memory network predicts that its failure risk will increase. The two prediction results are then integrated to generate a comprehensive equipment failure probability prediction value.

[0058] Resource demand classification and priority sorting, using the decision tree algorithm, divides the resource urgency level into high, medium, and low according to the set equipment failure probability threshold. Combined with the energy demand fluctuation forecast value, a dynamic weight table for resource scheduling is generated according to the given weight calculation formula. When the failure probability of a core device exceeds the threshold and energy demand is in short supply, the corresponding resource urgency level is high, and maintenance manpower, backup equipment, and sufficient energy are prioritized during scheduling.

[0059] A multi-agent reinforcement learning algorithm is used to generate scheduling strategies, define a global reward function, generate joint action strategies through a centralized training-distributed execution framework, and use policy distillation technology to migrate the global strategy to the local decision-making modules of each agent.

[0060] Define a global reward function. According to the park operation objectives, determine the global reward function in the form of a multi-objective weighted sum. At the current stage, the cost control pressure is high, and the weight of the cost-related reward will be set high. For example, for every 1% reduction in cost, the reward will increase by 10 points; when the energy efficiency is significantly improved, energy-related reward points will be given, such as a 5% reduction in energy consumption, the reward will increase by 8 points; if the carbon emissions are lower than the standard value, carbon emission-related rewards will be given in proportion. By dynamically adjusting the weight coefficient, the scheduling strategy is guided to develop in the optimal direction.

[0061] Centralized training-distributed execution framework. During the training phase, various intelligent agents in the park, such as the agent responsible for logistics scheduling, the energy management agent, and the equipment maintenance agent, share local observation information. For example, the logistics agent summarizes road congestion and vehicle en route location, while the energy agent summarizes energy reserves, real-time power generation, and other information. They all participate in training together to generate joint action strategies. For example, in the face of sudden order peaks and tight energy supply, the joint strategy coordinates logistics vehicles to prioritize the delivery of urgent orders, the energy agent to rationally allocate energy to ensure key production, and the equipment maintenance agent to inspect equipment prone to failure in advance.

[0062] Policy distillation technology is applied. After training is completed, the global strategy is migrated to the local decision-making module of each intelligent agent through policy distillation technology. Taking the logistics intelligent agent as an example, the optimal delivery path selection strategy under different road conditions and order requirements learned from the joint strategy is localized, so that it can make quick decisions based on local real-time information during actual operation without frequent interaction with other intelligent agents.

[0063] The robustness of the scheduling strategy is simulated and verified in the digital twin model, and the strategy is adjusted based on the simulation results. The final scheduling instructions are generated and sent to the park execution terminal.

[0064] Using the Monte Carlo simulation method, a large number of random disturbance tests are conducted on the generated scheduling strategy to simulate different equipment sudden failure scenarios, such as the sudden shutdown of core production equipment and large fluctuations in energy demand. For example, due to a power grid failure, the park needs to urgently allocate energy internally. The scheduling strategy is simulated multiple times to observe the changes in key indicators such as the overall production, logistics, and energy supply of the park, and evaluate its adaptability. If it is found that a certain strategy causes serious logistics congestion and production stagnation under specific failure scenarios, the problem will be marked in a timely manner.

[0065] The instructions are issued and monitored and adjusted. The final dispatch instructions after verification and optimization are transmitted in real time to various execution terminals in the park through the industrial Internet of Things platform, including factory automation production line control systems, logistics vehicle dispatch terminals, energy supply equipment control cabinets, etc. The execution terminals perform operations according to the instructions and monitor the execution of the dispatch instructions in real time through built-in feedback mechanisms, such as sensors sending back execution results and equipment operating status updates. If execution deviations are found, such as logistics vehicles deviating from the scheduled route or equipment failing to adjust power as required, the adjustment mechanism will be triggered immediately, and the control center will re-optimize the dispatch or remotely intervene to correct it.

[0066] In the step of using a graph neural network to perform relationship modeling, the nodes include equipment type, equipment status, energy consumption rate, and logistics throughput, and the edges include resource flow type, resource flow rate, and resource flow priority;

[0067] The attention mechanism determines the association weights between the nodes by calculating the Euclidean distance and cosine similarity between the nodes, and the association weights are used to adjust the edge weights of the graph neural network.

[0068] The step of performing resource demand analysis based on the multi-model prediction framework further includes:

[0069] Use long-short-term memory networks to model the time series data of equipment operating status to predict the potential trend of equipment failure. The prediction results are then integrated with the prediction results of the support vector machine algorithm to generate a comprehensive prediction value of equipment failure probability.

[0070] The decision tree algorithm’s priority sorting logic includes: dividing resource urgency levels according to equipment failure probability thresholds, and combining them with energy demand fluctuation forecasts to generate a dynamic weight table for resource scheduling. The weight calculation method is:

[0071] W=α·P fault +β·|D predicted-D current |

[0072] Among them, α, β are adjustable parameters, P fault is the failure probability, D predicted and D current They are forecast demand and current demand respectively.

[0073] The method of generating a joint action strategy through a centralized training-distributed execution framework and migrating the global strategy to the local decision-making module of each agent using policy distillation technology includes:

[0074] Define the global reward function as a weighted sum of multiple objectives:

[0075] R global =λ1·R cost +λ2·R energy +λ3·R carbon

[0076] Among them, λ1, λ2, λ3 are dynamically adjusted weight coefficients; R global is the global reward, R cost is the reward related to the cost, R energy Rewards related to energy efficiency, R carbon incentives tied to carbon emissions;

[0077] A centralized training-distributed execution framework is used to share local observation information of intelligent agents during the training phase and generate joint action strategies.

[0078] The global strategy is transferred to the local decision-making module of each agent through policy distillation technology.

[0079] The step of simulating and verifying the robustness of the scheduling strategy in the digital twin model further includes:

[0080] Conduct random perturbation tests on the scheduling strategy using Monte Carlo simulation methods to evaluate its adaptability to different failure scenarios and demand fluctuations, and optimize and adjust the scheduling strategy based on the test results;

[0081] The step of generating the final scheduling instruction and sending it to the park execution terminal also includes:

[0082] The scheduling instructions are transmitted to the execution terminal in real time through the industrial Internet of Things platform, and the execution of the scheduling instructions is monitored and adjusted in real time through the feedback mechanism of the execution terminal.

[0083] Example 2

[0084] Multi-objective collaborative optimization scheduling system for industrial parks based on artificial intelligence, such as Figure 2 Shown, including:

[0085] Data collection module, used to collect equipment operating status, energy consumption and logistics information within the industrial park using IoT technology;

[0086] Digital twin construction module, used to build digital twin models and dynamic topology maps based on collected data;

[0087] The relationship modeling module is used to perform relationship modeling using graph neural networks and output logistics path optimization solutions and energy allocation weights;

[0088] Resource demand analysis module, used to perform resource demand analysis based on a multi-model prediction framework;

[0089] Scheduling strategy generation module, used to generate scheduling strategies using multi-agent reinforcement learning algorithms;

[0090] The verification and adjustment module is used to simulate and verify the robustness of the scheduling strategy in the digital twin model, adjust the strategy according to the simulation results, and generate the final scheduling instructions.

[0091] An electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence are implemented.

[0092] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-mentioned artificial intelligence-based multi-objective collaborative optimization scheduling method for industrial parks.

[0093] In the specific implementation, various types of IoT sensors are deployed, and high-precision vibration and temperature sensors are installed on equipment such as placement machines and reflow soldering machines in the electronic manufacturing workshop to monitor the operating status of the equipment in real time, determine whether the equipment is operating normally and whether there are potential failure risks; energy consumption monitoring and operating time recording sensors are installed on ultra-low temperature refrigerators, centrifuges and other equipment in biomedical research and development laboratories to grasp the details of energy consumption; in terms of logistics, RFID readers and tags are fully laid at the entrances and exits of the park, warehouse shelves, and transport vehicles to accurately track the flow of raw materials, semi-finished products, and finished products. These sensors transmit data to the park data center through low-power Bluetooth or Wi-Fi networks.

[0094] Based on the massive amount of real-time data collected, the digital twin construction module uses advanced modeling tools to build a digital twin model of the park. Based on the park's geographic information, it presents the appearance and internal layout of each factory building, storage facilities, and energy supply station in three dimensions. For electronics factories, the model is detailed to the layout of equipment for each process of the chip packaging production line; biopharmaceutical factories accurately restore key areas such as cell culture rooms and purification workshops. At the same time, the dynamic topology map reflects the flow of resources in real time, such as the raw material transportation path from the chemical raw material warehouse to the new material synthesis workshop, and the power distribution link from the energy station to each high-energy-consuming workshop, and is updated in real time as the production progresses.

[0095] The relationship modeling module starts the graph neural network and defines node features including equipment process types, such as the SMT process in electronic manufacturing and the corresponding equipment for the fermentation process in biomedicine, equipment health status, energy consumption rate, and logistics processing capacity; edge features cover resource flow categories, flow rates, and priorities. By calculating the attention mechanism of Euclidean distance and cosine similarity between nodes, the connection between key nodes is strengthened. For example, high correlation weights are given to equipment chain nodes that ensure the production of core products. The final output is the optimized logistics route to avoid peak traffic areas in the park and shorten transportation time; energy is reasonably allocated to ensure stable power supply for key process links.

[0096] Using a multi-model prediction framework, the random forest model combines historical energy consumption data and weather forecasts. Temperature affects the heat dissipation energy consumption of electronic equipment, humidity is related to the energy consumption of biomedicine storage, and corporate production plans predict energy demand fluctuations, and plan energy allocation in advance. The support vector machine algorithm refers to the equipment's past failure data and real-time operating parameter changes to predict the probability of failure; the long-short-term memory network tracks the performance degradation trend of the equipment, and integrates the results of the two to accurately warn. The decision tree algorithm divides resource urgency according to the failure probability threshold (such as 0.7 is set as high risk), and combines energy demand forecasts to generate a dynamic weight table to prioritize the maintenance of high-risk equipment and the supply of key production energy.

[0097] Define a global reward function and adjust the weight according to the park development strategy. If the current focus is on green development, the carbon emission-related reward coefficient will be increased; if it is in the peak period of order delivery, the weight of production efficiency and cost control will be increased. A centralized training-distributed execution framework is adopted. Logistics, energy, equipment maintenance and other intelligent agents share local information such as workshop congestion, energy reserves, and equipment inspection results. Joint training generates strategies for dealing with complex scenarios. In the event of a sudden power failure, logistics is coordinated to suspend non-emergency delivery, the energy intelligent agent switches to backup power to protect key workshops, and the equipment maintenance intelligent agent investigates hidden dangers. The global strategy is then localized through policy distillation, and each intelligent agent makes quick decisions based on local real-time information.

[0098] In the digital twin model, Monte Carlo simulation is used to introduce disturbances such as random equipment failures, energy supply fluctuations, and temporary changes in orders. The scheduling strategy is tested multiple times, and indicators such as production continuity, energy waste, and logistics delays are observed to evaluate robustness. If a strategy is found to cause the production line to stop waiting for materials when multiple equipment fails, the material allocation and spare equipment activation rules are optimized in a targeted manner, and ultimately precise scheduling instructions are generated. These instructions are sent to the factory automation system, logistics scheduling terminal, and energy control cabinet through the Industrial Internet of Things platform to achieve real-time and efficient execution.

[0099] The park control center is equipped with high-performance electronic equipment. The processor adopts a multi-core architecture, has powerful computing capabilities, and can quickly process complex algorithms; the storage includes high-speed memory for temporarily storing real-time data, and a large-capacity hard disk stores historical data, model parameters and program code. After the computer program is loaded into the memory, it is executed by the processor and drives each module according to the established scheduling method and steps to ensure smooth operation of the system and real-time response to dynamic changes in the park.

[0100] A highly reliable solid-state drive is selected as the computer-readable storage medium to store the complete scheduling system program. It has fast reading and writing characteristics. Whether it is the system startup loading program or updating model parameters and optimization algorithms during operation, it can be completed efficiently, ensuring that even in the event of emergencies such as power outages and equipment restarts, the system can recover quickly and stably support the intelligent operation of the park.

[0101] The same or similar reference numerals correspond to the same or similar components;

[0102] The terms used in the drawings to describe positional relationships are for illustrative purposes only and should not be construed as limiting this patent;

[0103] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation methods here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. The multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence is characterized by: The following steps are involved: Build a digital twin model and generate a dynamic topology map. Leverage IoT technology to collect information about the operating status of equipment, energy consumption, and logistics within the industrial park. Based on the collected data, build a digital twin model and a dynamic topology map. The nodes of the dynamic topology map include factory, warehouse, and energy station entities, and edges represent resource flow relationships between nodes. Utilize graph neural networks for relationship modeling, define node and edge features, calculate the association weights between nodes through an attention mechanism, and use the graph neural network to learn the resource flow characteristics between nodes in the industrial park, outputting logistics path optimization solutions and energy allocation weights. Perform resource demand analysis based on a multi-model forecasting framework, using a random forest model to predict energy demand fluctuations, a support vector machine algorithm to predict equipment failure probability, and a decision tree algorithm to classify and prioritize resource demands; The step of performing resource demand analysis based on the multi-model prediction framework further includes: Use long-short-term memory networks to model the time series data of equipment operating status to predict the potential trend of equipment failure. The prediction results are then integrated with the prediction results of the support vector machine algorithm to generate a comprehensive prediction value of equipment failure probability. The decision tree algorithm’s priority sorting logic includes: dividing resource urgency levels according to equipment failure probability thresholds, and combining them with energy demand fluctuation forecasts to generate a dynamic weight table for resource scheduling. The weight calculation method is: W=α·P fault +β·|D predicted -D current | Among them, α, β are adjustable parameters, P fault is the failure probability, D predicted and D current They are forecast demand and current demand; A multi-agent reinforcement learning algorithm is used to generate scheduling strategies, define a global reward function, generate joint action strategies through a centralized training-distributed execution framework, and use policy distillation technology to migrate the global strategy to the local decision-making modules of each agent. The robustness of the scheduling strategy is verified by simulation in the digital twin model. The strategy is adjusted based on the simulation results, and the final scheduling instructions are generated and sent to the park execution terminal. The step of simulating and verifying the robustness of the scheduling strategy in the digital twin model further includes: Conduct random perturbation tests on the scheduling strategy using Monte Carlo simulation methods to evaluate its adaptability to different failure scenarios and demand fluctuations, and optimize and adjust the scheduling strategy based on the test results; The step of generating the final scheduling instruction and sending it to the park execution terminal also includes: The scheduling instructions are transmitted to the execution terminal in real time through the industrial Internet of Things platform, and the execution of the scheduling instructions is monitored and adjusted in real time through the feedback mechanism of the execution terminal.

2. The multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence according to claim 1 is characterized in that: In the step of using a graph neural network to perform relationship modeling, the nodes include equipment type, equipment status, energy consumption rate, and logistics throughput, and the edges include resource flow type, resource flow rate, and resource flow priority; The attention mechanism determines the association weights between the nodes by calculating the Euclidean distance and cosine similarity between the nodes, and the association weights are used to adjust the edge weights of the graph neural network.

3. The multi-objective collaborative optimization scheduling method for industrial parks based on artificial intelligence according to claim 1 is characterized in that: The method of generating a joint action strategy through a centralized training-distributed execution framework and migrating the global strategy to the local decision-making module of each agent using policy distillation technology includes: Define the global reward function as a weighted sum of multiple objectives: R global =λ1·R cost +λ2·R energy +λ3·R carbon Among them, λ1, λ2, λ3 are dynamically adjusted weight coefficients; R global is the global reward, R cost is the reward related to the cost, R energy Rewards related to energy efficiency, R carbon incentives tied to carbon emissions; A centralized training-distributed execution framework is used to share local observation information of intelligent agents during the training phase and generate joint action strategies. The global strategy is transferred to the local decision-making module of each agent through policy distillation technology.

4. An artificial intelligence-based multi-objective collaborative optimization scheduling system for industrial parks, implemented by the method described in any one of claims 1 to 3, characterized in that: include: Data collection module, used to collect equipment operating status, energy consumption and logistics information within the industrial park using IoT technology; Digital twin construction module, used to build digital twin models and dynamic topology maps based on collected data; The relationship modeling module is used to perform relationship modeling using graph neural networks and output logistics path optimization solutions and energy allocation weights; Resource demand analysis module, used to perform resource demand analysis based on a multi-model prediction framework; Scheduling strategy generation module, used to generate scheduling strategies using multi-agent reinforcement learning algorithms; The verification and adjustment module is used to simulate and verify the robustness of the scheduling strategy in the digital twin model, adjust the strategy according to the simulation results, and generate the final scheduling instructions.

5. An electronic device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method implements the steps of the multi-objective collaborative optimization scheduling method for an industrial park based on artificial intelligence as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the multi-objective collaborative optimization scheduling method for an industrial park based on artificial intelligence as described in any one of claims 1 to 3.

Images